The methodological worth of the Delphi forecasting technique

Focusing on the general issues of reliability and validity, several specific problems evident in reported applications of the Delphi method to forecasting are discussed. Reliability threats, which are noted in a number of procedural aspects of the method, arise from ill-considered procedural variations and lack of standardization. While validity threats are also found in several procedural aspects of the tool, they arise principally from pressures for convergence of predictions. This feature of the method, along with certain structural characteristics, is found to undermine critically its forecasting ability. Having discussed in some detail the nature of these difficulties, the paper closes with consideration of the reasons for the continued use of Delphi in spite of its shortcomings and with comments on alternative approaches.